Overview

For your second portfolio project, you’ll apply what you’ve learned about wrangling data. Your goal is to identify cities in the U.S. that struggle with weather prediction and explore possible reasons/patterns to why. Speciﬁcally, you will focus on the error in high and low temperature forecasting, and you may wish to also consider precipitation and outlook.

You should be careful about summarizing and joining data, and be on the lookout for data quality issues!

You should write a short report in a “blog post” format describing your ﬁndings, including an introductory paragraph that provides some context to your data and a couple paragraphs outlining your ﬁndings. That’s it. I’m looking for something that is insightful and well-crafted, rather than long and exhaustive.

You should write your blog post in quarto, create any graphics using ggplot2, and use tools from this class for data wrangling. To submit your work, push both your quarto file and rendered output document to GitHub and submit your pdf to Gradescope. Do not forget to give your post an informative title!

Data

The data for this portfolio problem is from the National Weather Service (with a very minor tweak from me, related to the criteria for earning an excellent on this assignment). The data includes sixteen months of forecasts and observations from 167 cities, as well as a separate data set with information about those cities and some other American cities.

Your repos will contain the following files:

data/cities.csv
data/outlook_meanings.csv
data/weather_forecasts.csv

`weather-forecasts.csv`

variable	class	description
date	date	date described by the forecast
city	factor	observation city
state	factor	state or territory
high_or_low	factor	whether the forecast is for the high temperature or the low temperature
forecast_hours_before	integer	the number of hours before the observation (one of 12, 24, 36, or 48)
observed_temp	integer	the actual observed temperature on that date (high or low)
forecast_temp	integer	the predicted temperature on that date (high or low)
observed_precip	double	the observed precipitation on that date, in inches; note that some observations lack an indication of precipitation, while others explicitly report 0
forecast_outlook	factor	an abbreviation for the general outlook, such as precipitation type
possible_error	factor	either (1) “none” if the row contains no potential errors or (2) thename of the variable that is the cause of the potential error

`cities.csv`

variable	class	description
city	character	city
state	character	state or territory
lon	double	longitude
lat	double	latitude
koppen	character	Köppen climate classiﬁcation
elevation	double	elevation in meters
distance_to_coast	double	distance_to_coast in miles
wind	double	mean wind speed
elevation_change_four	double	greatest elevation change in meters out of the four closest points to this city in a collection of elevations used by the team at Saint Louis University
elevation_change_eight	double	greatest elevation change in meters out of the eight closest points to this city in a collection of elevations used by the team at Saint Louis University
avg_annual_precip	double	average annual precipitation in inches

`outlook_meanings.csv`

variable	class
forecast_outlook	character
meaning	character

Submission

Your submission will be a short report detailing your findings.

Rubric

A successful project will:

Include all necessary components pushed to GitHub and pdf uploaded to Gradescope
Contain code, plots, and written summary in a single .qmd document
Include a .qmd with all necessary and no unnecessary code
Use both weather_forecasts.csv and cities.csv
Investigate the accuracy of the forecasts
Investigate possible reasons some parts of the country struggle with accuracy
Correctly import, join, and wrangle data sets
Correctly manipulate and use variable types (e.g. dates are treated as dates, factors are treated as factors)
Meet minimum submission quality standards
- Very few grammatical mistakes, spelling mistakes, or typos
- Informative title for your report is included
- Any graphs are readable with appropriate titles and labels
- The rendered document does not contain any unnecessary content (package loading messages, warnings, etc.)
- The rendered document does not echo your code

An excellent project will meet all of the requirements for a successful project, plus

Identify the coding/data quality issue within the cities.csv data set and use regex to correct it
- Include an appropriate visual illustrating some finding that uses the cell(s) or column(s)/variable(s) that you corrected using regex above
Meet high submission quality standards
- No grammatical mistakes, spelling mistakes, or typos
- Graphs have been customized (theme, color palette, scales, size, etc.)

Can I work with someone?

This project can be done individually or in pairs. If you’d like to work with a partner, please notify me by 11am on Friday of Week 5 so I can make a joint repo for you.